The Oxyrhynchus Papyri, Part 13

not vice versa. In v. 15 the employment of hagios holy by BA three times within the same sentence, referring to different persons in each case, is intolerable, and the addition of tas proseuchas ton hagion the prayers of the holy ones looks like a Christian gloss on prosanapherousin they offer up, which is intelligible by itself, while BA’s hagion angelon holy angels (angelon angels א; hagion holy ones 1594 and the third recension) may be the result of a conflation of readings or of a confusion between agion and aglon, a contraction of angelon angels found e.g. in 1603. 12. In v. 18 1594 has ego meth’ hymon ouch hoti te emautou chariti emen alla te thelesei tou theou I was with you, not that by my own grace, but by the will of God corresponding to B’s hoti ou te emautou chariti alla te thelesei tou theou hymon that not by my own grace but by the will of your God without a verb, which is supplied by A (add. elthon I came). The phrase ‘your God’ is very inappropriate in the mouth of an angel, and it is noticeable that the third recension, which at this point follows BA rather than א, ignores hymon your. The explanation is probably that hymon your had really nothing to do with theou God, but is the survival of emen meth’ hymon I was with you found in both 1594 and א, and that A’s elthon I came is merely a correction inserted to restore the defective grammar. 1594’s phrase ouch hoti . . . not that . . . in place of BA’s (hoti that) ouchi . . . not . . . gives a more literary touch to the passage, and might easily cause difficulty to some one who did not understand that emen I was was to be supplied with ego meth’ hymon I with you, with the result that a simpler construction was substituted. Fourthly, the result of an attempt to combine the merits of BA and א is partly extant in the third recension, and though that edition now appears to have taken into consideration the text represented by 1594 as well as those of א and BA (cf. p. 2); it does not coincide with 1594, and is in fact nearer to א than to 1594 or BA, just like 1076. That fragment on account of its affinity to א is still to be considered as probably a specimen of the missing portion of the third recension, not as part of the recension illustrated by 1594. We are therefore disposed to regard 1594 as an earlier form of the BA text, which developed out of 1594 partly owing to certain editorial changes, partly owing to corruptions introduced in the normal course of transmission.

There remains the question whether 1594 or א more closely represents the original text of Tobit. Owing to the small size of the fragment it is difficult to speak with certainty; but with regard to the characteristics of the BA text which Simpson (Journ. of Theol. Stud. xiv. 527–8) selects as evidence for the later date of BA it is noticeable that (1) 1594 does not tend, like BA, to avoid kai and as a connecting particle, (2) if 1594 is less redundant than א in ll. 14–18, in ll. 19–20 it has a repetition which is absent from א, and (3) the two uncommon words in 1594, prosanapherousi they offer up and optanomen I appeared, and the unusual construction in ll. 14–16 are absent from א, though as a rule the BA text is more commonplace than that of א. The א text is certainly not conspicuously better than that of 1594 in these six verses. The addition in א of Sarran Sarah before